SPEECHDAT server

- Last update 7 March, 1996 -

Welcome to the SPEECHDAT repository

The SpeechDat project is a CEC-funded initiative (LRE-63314) that addresses the fields of production, standardization, evaluation and dissemination of Spoken Language Resource (SLR).

The goal of the present repository is to provide guidelines and recommendations issued from the work achieved in SpeechDat (focusing on telephone like applications) as well as from other relevant current or past initiatives (covering other applications).
Among the deliverables of the project, recommendations will be available through different reports addressing:

Short and mid-term databases for applications

Database requirements for teleservices (March 95)
Existing Data Multimedia Catalogue (July 95)
Procedure for exchanging existing data (July 95)
Models of short/mid-term application directed database specification (July 95)
Specification of short/mid-term databases (July 95)
Reduced needs and specifications for the database (March 95)
Dissemination model and case study evaluation (Aug 95)

Resources for mid- and long-term spoken language research

Quality assessment and usability of existing corpora (May 95)
Corpora dissemination (Sept 95)
Delphi study (May 95)
Issues related to coverage within and over languages; recommendations for optimal sampling (Sept 95)
Inventory of resource requirements for future spoken language research (Sept 95)
Future applications of speech research (Sept 95)

Working standards, distribution and production of SLR

Working standards for short and medium application directed databases (March 95)
General working standards (Aug 95)
Recording tools (Aug 95)
Assessment tools (Aug 95)
Feasibility of automating phonetic annotation and building pronunciation lexica from corpus material (Sept 95)
Validation of databases (March 95)
Advanced data distribution (May 95)
Tasks of an ELRA (European Language Resources Association) (Jan 95)
Outline of the juridical, commercial and organisational form of an ELRA (Feb 95)
Relationship with on-going projects aiming at the dissemination of SLR (April 95)
Proposed organization mechanism for production of future SLR (Aug 95)

This list will be regularly updated with the availability of the reports as long as the project progresses.

The SpeechDat project

One of his main tasks is to help to create an European Infrastructure for distribution and evalution of Language Resources. Whereas the MLAP-projects PAROLE and POINTER are responsible to deal with written language resources and terminology, SpeechDat is in charge to concentrate on SLR. Due to the progress made in all these preparatory projects with the help of the project RELATOR, a European Language Resource Association (ELRA) has been founded. Within ELRA Speechdat is in charge to build up a validation center for SLR, which is currently SPEX, to set standards for SLR as they are needed within the SpeechDat project and to put in the distribution center of ELRA first SLR. The working standards and the validation procedures as described in the deliverables of Speechdat, should give some guidelines for the running LRE-Call for producing SLR.

Contact points

CEC officer: Jose Soler
COMMISSION OF THE EUROPEAN COMMUNITIES
DIRECTORATE GENERAL
Telecommunications, information market and exploitation of research
Information industry and market and language processing
Basic studies in linguistics
Luxembourg
XIII - E5
+352 4301 34655 +352 4301 33291

More about SpeechDat ?

Related sites

Other valuable sources of information about these topics exist in the:

repository of the Relator project (LRE 62-056 )
FTP site of the EAGLES project (LRE 61-100)
FTP site of the US Linguistic Data Consortium

© SpeechDat
This server is maintained at Institut de la Communication Parlée, Grenoble, France.
Comments are welcome: zeiliger@icp.grenet.fr